A Coalition Takes Shape

On Monday, Nvidia and more than 30 technology, cybersecurity, and enterprise software heavyweights announced the formation of the Open Secure AI Alliance. The goal? Build and share open source tools, models, and techniques for securing AI systems and agents.

The alliance builds on earlier work from the Linux Foundation’s Akrites initiative and the OpenSSF community. It’s a recognition that AI security can’t stay locked inside proprietary labs.

Who’s In

The inaugural partner list reads like a who’s who of tech. Adobe, Cisco, Cloudflare, CrowdStrike, Databricks, Dell, IBM, Microsoft, Palo Alto Networks, Red Hat, Salesforce, SAP, Snowflake, and Hugging Face are all in. So are Capital One, Cloudera, Cognition, DoorDash, Elastic, HPE, LangChain, Naver, NetApp, Nous Research, OpenClaw, Palantir, Reflection AI, ServiceNow, Siemens, SK Telecom, SpaceXAI, Synopsys, Thinking Machines Lab, and TrendAI.

That’s a broad tent. Banks, cloud providers, chipmakers, security vendors, and AI startups — all agreeing on one thing: defenders need better tools.

What Each Giant Is Bringing

Nvidia’s contribution includes open models, weights, data, and agent harness research. The company also released a new open source project called NOOA, designed to help harnesses make agent behavior easier to trace, test, and audit.

HPE is contributing to SPIFFE/SPIRE, a zero-trust identity framework for cryptographically verifying AI agents and services. Hugging Face is donating its Safetensors model weight storage format to the PyTorch Foundation.

IBM and Red Hat are extending open source supply chain security through the Lightwell project, which aims to deliver automated vulnerability remediation at scale. Microsoft is contributing MDASH, a multi-model agentic scanning harness that coordinates AI agents to find, debate, and validate exploitable software bugs. SpaceXAI is open-sourcing its Grok Build terminal-based AI coding agent, with plans to eventually open-source the weights of the Grok model line.

Why Openness Matters Now

The alliance’s core argument: open models, harnesses, and security tooling should be treated as defensive assets, not liabilities. The group warns policymakers and regulators that broad restrictions on open frontier AI could weaken collective cyber defense capacity.

Nvidia points to a recent security incident involving OpenAI and Hugging Face. When closed AI tools couldn’t differentiate between attackers and defenders and blocked forensic work, Hugging Face used the open-weight GLM 5.2 model on its own systems to review over 17,000 actions and contain the breach.

“The right response is not to deny defenders access to capable open systems,” Nvidia said. “It is to pair openness with strong safeguards, clear rules against malicious misuse, rigorous evaluation and rapid remediation. In cybersecurity, the safer path is the one that gives more defenders the ability to test, verify and strengthen the systems on which society relies.”

“Defenders need both frontier closed models and frontier open models, working together,” the company added, “so they can choose the right system for the job and ensure that transparency, adaptation and sovereign control are available wherever security demands them.”

What This Means for Security Teams

For practitioners, the alliance could mean more options when it comes to AI security testing. Open harnesses like NOOA and MDASH give teams the ability to probe their own AI systems without waiting for a vendor’s patch cycle.

The zero-trust identity work from HPE on SPIFFE/SPIRE also matters. As AI agents become more autonomous, verifying who — or what — is making decisions becomes critical. Cryptographic identity for agents isn’t a nice-to-have anymore.

The supply chain angle is worth watching too. Lightwell’s automated vulnerability remediation could help close the gap between discovery and patching, a problem that’s only getting worse as AI-generated code accelerates development cycles.

The Bigger Picture

This launch lands amid a broader push on AI security. The White House recently launched an AI-driven vulnerability coordination initiative called ‘Gold Eagle.’ Meanwhile, researchers keep finding flaws in AI systems — from ChatGPT agent vulnerabilities to nuclear-sabotage malware benchmarks that trip up frontier models.

The AI security alliance is a bet that openness will win. It’s also a direct challenge to the idea that keeping powerful AI models locked away makes the world safer. The alliance’s answer: locked models leave defenders blind.

Whether that argument sways regulators remains to be seen. But for now, the tools are coming. Open source, auditable, and ready for defenders to use.

The New Face of Contagious Interview

The Contagious Interview campaign has been a thorn in the side of cybersecurity teams for months. But its latest evolution is raising alarms. North Korean threat actors have added a new trick: hiding malicious payloads inside SVG image files using steganography.

Security researchers recently uncovered a wave of attacks where fake job postings lure developers into running coding challenges. Those challenges, it turns out, are Trojan horses. Any user who ran the project ended up with a four-stage payload aligned with OtterCookie malware — a browser credential and crypto wallet stealer, a file stealer, and more.

How the SVG Steganography Attack Works

Steganography isn’t new. But using SVG files — the vector image format widely supported in browsers — is a clever twist. The malicious code hides in plain sight within the image’s XML structure, making it nearly invisible to traditional scanners.

The attack chain is deceptively simple:

• Victims receive a fake coding challenge via a job interview invitation.
• The project includes an SVG image that appears harmless.
• When the code runs, it extracts hidden payloads from the SVG’s metadata.
• The payload installs OtterCookie, which steals credentials, crypto wallet data, and files.

This multi-stage approach is designed to evade detection. Each step is small and innocuous on its own, but together they form a devastating pipeline.

Why SVG Files Are the Perfect Cover

SVG files are XML-based, which means they can contain arbitrary data in comments, attributes, or even within the image paths themselves. Security tools that scan for malware often overlook these files because they’re not executable by default.

Attackers exploit this blind spot. They embed encrypted blobs or base64-encoded strings inside the SVG, then use a script in the coding challenge to decode and execute them. The result? A stealthy delivery mechanism that slips past most defenses.

This isn’t just a technical curiosity. It’s a practical threat for anyone in tech, especially developers who are actively job hunting.

OtterCookie: What It Steals and How

OtterCookie is a modular stealer that targets:

• Browser credentials, including saved passwords and cookies
• Crypto wallet extensions and private keys
• Files matching specific extensions, like documents and source code
• System information for further exploitation

The malware communicates with a command-and-control server, exfiltrating data in real time. Researchers note that it’s often paired with other tools, making it a versatile component in a larger arsenal.

For victims, the damage can be severe. Stolen crypto wallets can be drained within minutes. Corporate credentials can lead to network intrusions and data breaches.

How to Protect Yourself from Fake Job Scams

If you’re a developer, this campaign is a direct threat. Here’s how to stay safe:

1. Verify the recruiter. Check the company’s official domain and email format. Be wary of unsolicited interview invitations.
2. Inspect the code. Before running any coding challenge, review the project files. Look for suspicious scripts, obfuscated code, or unusual image files.
3. Run in a sandbox. Execute unknown code in a virtual machine or isolated environment to contain any potential damage.
4. Monitor your accounts. Regularly check your browser extensions and crypto wallets for unauthorized access.

These steps won’t guarantee safety, but they drastically reduce your risk. The threat landscape is evolving, and so must your defenses.

The Bigger Picture: North Korea’s Cyber Operations

This campaign is part of a broader pattern. North Korean state-sponsored hackers have long used social engineering to infiltrate tech companies, steal intellectual property, and generate revenue through crypto theft. The Contagious Interview campaign is just one example of their relentless innovation.

Security researchers continue to track these groups, but the challenge is immense. As defenses improve, attackers adapt. SVG steganography is the latest adaptation, and it won’t be the last.

Staying informed is your best weapon. Follow cybersecurity news, update your tools, and always question what’s in front of you. A single suspicious file could be the difference between a secure career and a devastating breach.

Emergency Action Required for Federal Agencies

The U.S. Cybersecurity and Infrastructure Security Agency (CISA) moved quickly on Thursday, adding a freshly patched Microsoft SharePoint Server flaw to its Known Exploited Vulnerabilities (KEV) catalog. The designation means Federal Civilian Executive Branch (FCEB) agencies have a hard deadline: apply the fixes by July 19, 2026.

This isn’t a theoretical risk. The vulnerability, tracked as CVE-2026-58644, carries a CVSS score of 9.8 — critical, with a capital C. It’s a deserialization bug that attackers are already exploiting in the wild.

What Makes CVE-2026-58644 So Dangerous?

Deserialization flaws are nasty. They let an attacker manipulate serialized data passed to the application, potentially executing arbitrary code on the server. For SharePoint, that’s a crown jewel target — the server often sits deep inside corporate networks with broad access to files, credentials, and other systems.

The CVSS score of 9.8 reflects the severity: no user interaction required, low attack complexity, and the potential for full compromise of the affected server. Think about what a successful exploit means. An attacker gains a foothold, then moves laterally. SharePoint is frequently the starting point for ransomware gangs and nation-state actors alike.

Why the Zero-Day Label?

Microsoft patched the flaw before CISA’s announcement, but the fact that it was exploited before a fix existed makes it a zero-day. The window between discovery and patch is when the damage happens. CISA’s KEV listing confirms that window was used — and likely continues to be used by threat actors targeting unpatched systems.

What Federal Agencies Must Do Now

The binding operational directive (BOD) 22-01 requires FCEB agencies to remediate KEV-listed vulnerabilities by the specified due date. For CVE-2026-58644, that’s July 19, 2026. Miss it, and you’re in violation of federal policy. But the real cost isn’t bureaucratic — it’s the risk of a breach.

Beyond the federal sphere, CISA’s KEV catalog has become a de facto checklist for security teams everywhere. If you run Microsoft SharePoint Server, consider this your wake-up call. Patch immediately. Don’t wait for the 19th.

Practical Steps for SharePoint Administrators

Here’s what you should do today, not next week:

• Apply the latest Microsoft security update — the patch for CVE-2026-58644 is included in the June 2026 Patch Tuesday release. Verify your installation is current.
• Check your logs — look for unusual deserialization activity or unexpected process executions on SharePoint servers. Indicators of compromise may exist if you were targeted before patching.
• Harden your environment — restrict network access to SharePoint servers, enforce least-privilege accounts, and monitor for anomalous behavior.
• Review CISA’s KEV catalog regularly — it’s updated frequently. Make it part of your weekly threat intelligence routine.

Broader Implications for Enterprise Security

This isn’t an isolated incident. SharePoint has been a recurring target — remember the ProxyLogon and ProxyShell issues in Exchange? The pattern is consistent: on-premises collaboration servers are high-value targets. They hold data, they’re often internet-facing, and they’re complex to patch.

If you’re running SharePoint Server on-premises, the message is clear. You’re in the crosshairs. The shift to cloud-based services like SharePoint Online might reduce some risk, but hybrid deployments still carry on-premises components that need attention.

For security teams, the KEV catalog is a gift. It tells you exactly what’s being exploited right now. The hard part is acting on it. Prioritize CVE-2026-58644. Schedule the maintenance window. Communicate the urgency to stakeholders.

One more thing: don’t assume your vulnerability scanner caught this. Deserialization bugs can be tricky to detect. Manual verification of the SharePoint build number is a good sanity check.

The bottom line? This is a critical, actively exploited vulnerability with a federal deadline. Treat it with the seriousness it deserves. Patch. Verify. Monitor. That’s the playbook.